Implement better typing in EpicsName PandAName PviName

[evalott100]

Currently we define types for each of the above names:

PandABlocks-ioc/src/pandablocks_ioc/_types.py

Lines 20 to 51 in d949bd6

	# EPICS format, i.e. ":" dividers
	EpicsName = NewType("EpicsName", str)
	# PandA format, i.e. "." dividers
	PandAName = NewType("PandAName", str)
	# No dividers and PascalCase
	PviName = NewType("PviName", str)
	def panda_to_epics_name(field_name: PandAName) -> EpicsName:
	"""Convert PandA naming convention to EPICS convention. This module defaults to
	EPICS names internally, only converting back to PandA names when necessary."""
	return EpicsName(field_name.replace(".", ":"))
	def epics_to_panda_name(field_name: EpicsName) -> PandAName:
	"""Convert EPICS naming convention to PandA convention. This module defaults to
	EPICS names internally, only converting back to PandA names when necessary."""
	return PandAName(field_name.replace(":", "."))
	def epics_to_pvi_name(field_name: EpicsName) -> PviName:
	"""Converts EPICS naming convention to PVI naming convention.
	For example PANDA:PCAP:TRIG_EDGE -> TrigEdge."""
	relevant_section = field_name.split(":")[-1]
	words = relevant_section.replace("-", "_").split("_")
	capitalised_word = "".join(word.capitalize() for word in words)
	# We don't want to allow any non-alphanumeric characters.
	formatted_word = re.search(r"[A-Za-z0-9]+", capitalised_word)
	assert formatted_word
	return PviName(formatted_word.group())

But when used at runtime we don't have restrictions, and everything is a str.

We should turn these types into pydantic BaseModel and validate each against a regex schema on initialisation.